o_18ufhmfmq19t513t3lgmn5l1qa8a.pdf

292 CHAPTER 13 ■ DATABASE SUPPORT
link under the heading “Abbreviated.” You should now get a zip file containing a text file named
ABBREV.txt, along with a PDF file describing its contents. 1 If you have trouble finding this particular
file, any old data will do, of course. Just modify the source code to suit.
The data in the ABBREV.txt file has one data record per line, with the fields separated by
caret (^) characters. The numeric fields contain numbers directly, while the textual fields have
their string values “quoted” with a tilde (~) on each side. Here is a sample line, with parts deleted
for brevity:
~07276~^~HORMEL SPAM ... PORK W/ HAM MINCED CND~^ ... ^~1 serving~^^~~^0
Parsing such a line into individual fields is a simple as using line.split('^'). If a field
starts with a tilde, you know it’s a string and can use field.strip('~') to get its contents. For
the other (numeric) fields, float(field) should do the trick, except, of course, when the field is
empty. The program developed in the following sections will transfer the data in this ASCII file
into your SQL database, and let you perform some (semi-)interesting queries on them.
■Note This example program is intentionally simple. For a slightly more advanced example of database use
in Python, see Chapter 26.
Creating and Populating Tables
To actually create the tables of the database, and to populate them, writing a completely separate
one-shot program is probably the easiest solution. You can run this program once, and then
forget about both it and the original data source (the ABBREV.txt file), although keeping them
around is probably a good idea.
The program shown in Listing 13-1 creates a table called food with some appropriate fields,
reads the file ABBREV.txt, parses it (by splitting the lines and converting the individual fields using
a utility function, convert), and inserts values read from the text field into the database using
an SQL INSERT statement in a call to curs.execute.
■Note It would have been possible to use curs.executemany, supplying a list of all the rows extracted
from the data file. This would have given a minor speedup in this case, but might have given a more substantial
speedup if a networked client-server SQL system were used.
1. At the time of writing, you can get this file from the URL http://www.nal.usda.gov/fnic/foodcomp/
Data/SR17/dnload/sr17abbr.zip.